🩺 MedAI: Intelligent Medical Triage & Mapping

A Next.js 15 Full-Stack Application powered by Groq's Llama 3 models for instant, multimodal health analysis.

---

🚀 Overview

MedAI is an advanced, portfolio-grade medical guidance application. It leverages the cutting-edge Groq API (Llama 3.3 70B & Llama 3.2 11B Vision) to analyze user-reported symptoms and photos, outputting structured medical likelihoods and OTC recommendations.

The application utilizes a proprietary Agentic Triage Loop—making up to 3 rapid API calls to interrogate the user on their symptoms to mathematically increase its diagnostic confidence before generating a final PDF report.

Disclaimer: This is a software demonstration meant for informational purposes. It is not a substitute for clinical medical advice.

---

🛠️ Tech Stack & Architecture

Core Frontend

• Framework: Next.js 16 (App Router, Server Actions)
• UI & Styling: React 19, Tailwind CSS v4, Lucide React
• Mapping: Leaflet + MapTiler (Lazy-loaded via next/dynamic to drastically reduce initial JS payload)

Backend & AI Layer

• Database: Supabase PostgreSQL (utilizing strict Row Level Security)
• Auth: Supabase Auth (Magic Links / OAuth)
• AI Brain: groq-sdk
  • Text Router: Uses llama-3.3-70b-versatile for lightning-fast NLP processing.
  • Vision Router: Automatically falls back to llama-3.2-11b-vision-preview if a symptom photo is uploaded via native Base64 encoding.

---

✨ Key Features

• Agentic Multi-Turn Triage: Instead of a "one-shot" guess, the AI operates in a smart loop, asking follow-up questions to constrain hallucinatory diagnoses.
• Dynamic Female Health Context: Explicit algorithms compute menstrual phases based on cycle records to feed specific biological context into the AI engine.
• Drug Interaction Guardian: Implicitly cross-references the user's logged medications_list before suggesting any OTC treatments.
• Local Clinic Mapping: Renders interactive local maps directly in the UI for users to find the nearest physical doctors.
• Multilingual Web Speech API: Native microphone dictation and translation for English, Hindi, and Spanish speakers.
• Client-Side PDF Generation: Exports cleanly-styled, stamped PDF records of the AI's findings.

---

💻 Getting Started Locally

1. Clone & Install

   git clone https://github.com/Nikky-SDE/MedAI.git
   cd MedAI/web
   npm install

2. Environment Variables Create a .env.local file inside the web directory:

   NEXT_PUBLIC_SUPABASE_URL=your_supabase_url
   NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key
   NEXT_PUBLIC_MAPTILER_KEY=your_maptiler_key
   GROQ_API_KEY=your_groq_api_key

📐 PRD Reference

For a complete, architectural breakdown of the components designed in this application, please refer to the PRD.md file located in the repository root.